Supporting Sandtray Therapy on an Interactive

Transcription

Supporting Sandtray Therapy on an Interactive
Supporting Sandtray Therapy on an Interactive Tabletop
Mark Hancock1 , Thomas ten Cate1,2 , Sheelagh Carpendale1 , Tobias Isenberg2
2
1
University of Groningen, The Netherlands
University of Calgary, Canada
Institute of Mathematics & Computer Science
Department of Computer Science
[email protected],[email protected]
{msh,sheelagh}@cpsc.ucalgary.ca
ABSTRACT
We present the iterative design of a virtual sandtray application for a tabletop display. The purpose of our prototype is
to support sandtray therapy, a form of art therapy typically
used for younger clients. A significant aspect of this therapy
is the insight gained by the therapist as they observe the
client interact with the figurines they use to create a scene
in the sandtray. In this manner, the therapist can gain increased understanding of the client’s psyche. We worked
with three sandtray therapists throughout the evolution of
our prototype. We describe the details of the three phases
of this design process: initial face-to-face meetings, iterative
design and development via distance collaboration, and a
final face-to-face feedback session. This process revealed
that our prototype was sufficient for therapists to gain insight
about a person’s psyche through their interactions with the
virtual sandtray.
ACM Classification Keywords
H.5.3 Information interfaces and presentation: Group and
Organization Interfaces—CSCW.
General Terms
Design, Human Factors
INTRODUCTION
Social workers and therapists are developing new ways of
reaching and working with children who are troubled or have
experienced traumatic events and difficult life situations.
One such method is sandtray therapy [3, 7, 16]—a type of
play or art therapy [5] in which the use of figurines in a tray
of sand is a vehicle for establishing interaction and rapport
between the therapist and the child (Figure 1). Children
placing and moving the figurines in the sandtray provides a
venue by which therapists can observe the manner in which
the child thinks about their experiences and feelings. In
response to an idea from a practicing sandtray therapist, we
developed a virtual sandtray. In this paper we present its
iterative evolution in collaboration with three therapists who
make use of sandtray in their therapeutic practices.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee.
CHI 2010, April 1015, 2010, Atlanta, Georgia, USA.
Copyright 2010 ACM 978-1-60558-929-9/10/04...$10.00.
Figure 1: A sandtray typically used for sandtray therapy. Sandtray
therapists typically observe clients creating a scene or “playing” in the
sand to gain insight into their psyche (courtesy of Kristina Walter [29]).
There were many reasons for the request that triggered this
research. Sandtray therapy, often considered a type of art
therapy [5] because the therapy session involves the clients
creating a scene out of available supplies, has particular characteristics which make it well suited to an interactive tabletop. These include factors of age stereotyping, the characteristics of the sand itself, and the types of interaction that are
therapeutically beneficial. In terms of age stereotyping, the
associations of sandtrays (or sandboxes) are with activities
we did as young children. While this works well in establishing rapport with young children, it can pose problems
with young teenagers and pre-teens to whom activities in a
sandbox may seem just too uncool. This age group (10 to
about 13 or 14) is a particularly difficult age group for therapists to reach and is also a particularly sensitive age group
since so much is changing in their lives at these ages. One of
the therapists we are working with suggested that the ‘wow’
factor of a digital table might prove a great bridging factor.
Also, some children, perhaps due to their own response to
traumas, dislike the feeling of sand and refuse to play with it
and may find a digital sandbox more to their liking.
Perhaps most important is the potential for interaction;
sandtrays offer special interaction advantages. For example,
other forms of media used in art therapy [5], such as pencils,
paints, and clay, result in the client creating a ‘snapshot’
(e. g., pencil drawings, paintings, sculptures) as an end
result. In contrast, the temporary nature of the sand invites
play and, therefore, the creation of a narrative. The process
of creating a ‘scene’ containing several objects, in which
the narrative can unfold, can be particularly informative to
the therapist, who can often infer self-representation in one
object and, from there, the relationships to other objects in
the scene. Thus, the client’s interaction with the objects is of
particular importance to the therapist, and the possibilities
of multi-touch interaction make this application particularly
suitable to adoption with tabletop display technology. The
direct nature of touch input to tabletop displays affords
observation of these interactions, and the fact that the display
is digital makes the scene being created similarly temporary.
While this is an interesting application for tabletops, there
are also particularly interesting and challenging research issues. The therapists we worked with explained how in the
physical world they have become sensitive over time to understanding the possible implications of the active (manipulation of physical objects) part of the sandtray work. We
were particularly interested in whether this professional skill
would hold for virtual 3D interaction. In particular, is the
virtual medium rich enough for a scene to be constructed that
the therapists can understand, or are viewing issues, such as
the need to project onto a 2D screen, too limiting? Can a
virtual object take on a variety of meanings so as to enable
the telling of a story to the therapists, or will they be interpreted as mere data or information? Are the interactions on
a virtual table rich enough to convey meaning to the therapist about a client’s psyche, or is the disconnect between a
person’s actions and the surface’s reaction too great? Our
more general goal was to discover whether the therapists felt
that they could effectively perform therapy with this digital
sandtray or some future design iteration.
used in a game form to allow four children simultaneously
to train social skills and collaboration. Similar to our own
motivation, the authors name the children’s affinity to technology as one of the criteria that makes tabletop technology
well suited for such therapeutic applications.
Although not used to tell a story, Piper and Hollan [23] describe the design of an interactive table used to facilitate
communication between a doctor and a deaf patient. This design process has many similarities to our own and shows the
benefit of tabletop technology for the deaf community. As
would be expected, our work reinforces the idea that working
closely with the people who professionally understand the
application area can lead to successful tabletop applications.
The focus of our research differs from these storytelling and
therapeutic tabletop applications. We focus on the therapists
and their ability to interpret the actions being carried out on
the tabletop display itself. We leverage existing literature
that suggests the usability of our interaction techniques and
ask whether these techniques are rich enough to be interpreted by the observing therapist.
2D Tabletop Display Interaction
Storytelling and Therapeutic Tabletop Applications
Kruger et al. [17] showed that people use the orientation of
artifacts on a table to communicate with one another. This
form of silent communication is likely highly relevant to the
therapist-client communication process in sandtray therapy,
as the orientation of objects in the sandbox can be a key
element of interpreting a client’s thoughts. Their rotate n’
translate (RNT) technique [18] uses only one point of contact to enable this rich communication. Techniques such
as TNT [21] have extended this technique by mapping the
rotation of a hand or stylus to the orientation of the virtual
object. As modern touch technologies (e. g., [8, 11]) provide
more simultaneous touches, these additional contact points
can also be used to specify object position, orientation, and
scale. For example, a technique now commonly used in
multi-touch consumer appliances employs two points of contact to specify position, orientation, and size of objects [13].
Depending on the hardware that recognizes the input and
on the task to be performed, however, the suitability of the
different interaction techniques may vary [13].
Our specific application domain, virtual sandtray therapy, is
related to a number of approaches where modern touch and
tangible technology is used to support work with children for
storytelling or therapeutic purposes. Early examples include
the use of robotic stuffed animals [2] to help young cardiac
patients cope with their situation by encouraging them to
talk about it, comparable to virtual sandtray therapy. Later
work employed an interactive physical play mat (StoryMat)
to record children’s storytelling activities [6, 24]. Li et al.
[20] developed a tangible tabletop game to support the therapy of children with cerebral palsy who need to train specific
motor skills. The game combined tangible elements with a
tabletop surface that was illuminated with coloured LEDs
from below and was found to encourage children to train the
desired therapeutic movements. Morris et al. [22] describe
an interface designed for children with Asperger’s syndrome.
Here, the fact that tabletop displays afford collaboration is
In addition to applying simple transformations to objects,
it is often necessary to invoke other actions. While techniques known from traditional interfaces such as menus, tool
palettes, and buttons can be used for this purpose, touch
interfaces often also employ gestures. For example, Wu
and Balakrishnan [32] demonstrated how to control a room
planning application using a variety of gestures and hand
postures. Other approaches also combine posture and gesture interaction with speech input [28]. Even with relatively
simple size-based recognition hardware it is possible to use
a set of hand postures to parameterize or control actions
in an interface [10]. Input techniques that provide more
information on the shape of a touch, however, can be used to
define postures inspired by physical interaction and to infer
forces to be used in the interaction [4]. An approximated
touching force can also be used to control the layering of 2D
We present the cooperative design of a virtual sandtray
through three phases of design: initial face-to-face meetings,
iterative remote collaborative design, and a face-to-face
feedback session. We end with an in-depth discussion of
the results of this collaboration and a description of how to
make use of our findings beyond the digital sandtray.
RELATED WORK
In addition to the field of traditional sandtray therapy [3, 5,
7, 16], we review relevant research in the following areas:
technologies that support storytelling or therapy, and tabletop interaction with both 2D and 3D virtual objects.
objects [2], which is related to the concept of shallow-depth
3D interaction [12] on which our work is largely based.
We leverage some of this work by providing interaction techniques that allow the therapist to interpret the actions of the
client. In our system, we opted for interaction techniques
that explicitly convey the consequence of the action, as opposed to abstract gestures that must be learned. We also
chose to use physically familiar feedback from the system
as opposed to (for example) drastic global changes caused
by subtle actions, such as a button press.
3D Tabletop Display Interaction
While interaction techniques that rely on input captured in
3D space have been investigated in detail in 3D stereoscopic
environments such as the Responsive Workbench [19], we
concentrate on 2D input on the surface of a horizontal display to interact with a 3D scene. In particular, we use sticky
fingers and opposable thumbs [14], a technique based largely
on shallow-depth 3D techniques [12], to provide six degrees
of freedom (DOF) control of 3D objects. This shallow-depth
technique has been previously evaluated, demonstrating that
three touches can be both faster and more well-liked than
one or two touches. Another metaphor to translate 2D input to interaction with 3D shapes in a horizontal interface,
BumpTop, was introduced by Agarawala and Balakrishnan
[1]. This interface is based on a physics simulation with 3D
shapes, controlled through gestures as well as some menus.
A similar interface that also relies on physical simulation
was introduced by Wilson et al. [31] who extend the interacting objects such as fingers virtually into a 3D environment
in which objects are located. This system was also evaluated
and shown to be faster and well-liked by participants.
We leverage these 3D interaction techniques together with
the use of a physics engine to provide tabletop interaction
that is interpretable by the therapist. While these techniques
and their corresponding studies show the usability of 3D interaction on a table, our work shows that these same interaction techniques can also benefit the therapist observing these
interactions. By striking a balance between the ability to
precisely control the 3D object and having on-screen objects
react physically to those precise actions, the actions of the
client become rich enough for the therapist to interpret the
‘language’ of those actions from the client.
METHODOLOGY
We used a cooperative design process [9,26] with three sandtray therapists, who were involved throughout this research
to provide us with expert domain knowledge. There were
three phases of this design process: Phase I involved initial face-to-face meetings with one sandtray therapist (T1),
Phase II was an iterative distance collaboration (via phone
and email) with two therapists (T1 & T2) throughout the
design and implementation of the prototype, and Phase III
was a face-to-face discussion with two therapists (T1 & T3)
to provide feedback about the working prototype. The first
phase involved two meetings, a pre-planning meeting and
a follow-up meeting. At the former meeting, the idea of
implementing a digital sandtray had not yet been conceived.
This meeting was a demonstration of existing technology to
T1, which triggered a discussion about how a digital tabletop
sandtray might provide a solution to some of the comfort
problems some clients feel with traditional sandtrays. The
second meeting was to plan how we could collaboratively
design our prototype at a distance, as the therapists were
in a city approximately 630 km from our research lab. The
second phase in our design process involved extensive discussion via phone and email with both T1 and T2, who described in detail what they felt were the essential elements of
sandtray therapy. We include direct quotes from our email
communication with T2 throughout the next section (Phase I
& II: Designing the Virtual Sandtray). We responded to this
communication with descriptions of design ideas and questions about what made the elements important. These conversations were highly iterative and led to the design considerations that are described next. During design and development of the prototype, we maintained contact with both T1
and T2 to iterate and refine the design. When the prototype
was finished, T1 and T3 joined us in a day-long workshop to
use our prototype first-hand and provide feedback (Phase III:
Face-to-Face Feedback Session).
Our in-depth discussions with the therapists revealed that it
is common for them to constantly be collecting artifacts to
use in therapy sessions from the environment (e. g., sticks,
leaves, plastic cups, etc.). For example, one could consider
our first meeting with T1 to be an example of her ‘collecting’
our technology. More generally, their practice frequently
involves the evaluation of the suitability of tools and techniques for use in therapy. This skill is learned over time
and does not directly involve their clients. We thus focused
our research on the therapists; our research asks whether, in
the virtual world, therapists can still interpret a person’s actions in a meaningful way to perform therapy. We therefore
decided to only include therapists (and not clients) in our
design process. This decision has the consequence that our
findings should be interpreted with a therapist-focused lens.
We did, however, include a mock therapy session in our daylong workshop, where one of the designers played the role
of a client. Indeed, part of the training for students learning
to do sandtray therapy involves participating in a session as
a client themselves. The therapists in our study explained
that these mock therapy sessions were necessary in order
to better understand the experience for the client. Thus, it
seemed particularly appropriate for our own understanding
to undergo a similar experience.We rely on previous studies
[12,31] to validate the usability of the interaction techniques
used by the client.
PHASE I & II: DESIGNING THE VIRTUAL SANDTRAY
There are two possible avenues to explore for features in a
digital sandtray: those already offered by a physical sandtray,
perhaps adapted for use on a tabletop, and new options that
do not exist in the physical world, but are made possible by
the virtual. Our list of design considerations (DC1–DC4)
contains some features from traditional sandtrays that the
therapists thought were important to maintain and some digital features they thought it would be interesting to explore.
In our bottom-up approach we started with nothing, adding
library of 3D models, around 160 figurines of many different
categories were selected for use in the virtual sandtray.
“With a sand tray it is rare that a client will ask for a specific object, precisely because inspiration tends to start
with associations to the presented repertoire of objects.
This puts the client in the position of immediately symbolizing and associating. While the (relatively) fixed nature of the objects limits the expressiveness of the work,
the way that they are animated and placed becomes
the client’s means of articulating their own meanings
regarding those objects.”
–T2
However, the sand itself can be used as an open medium that
can be shaped at will and made into a backdrop for the story.
Because directly simulating the behaviour of sand is computationally intensive, we decided to provide a different type
of background open media in textured ‘paint’. The specific
types of paint we included represent different surfaces, such
as sand, grass, concrete and water.
Figure 2: A shelf of figurines used for sandtray therapy (courtesy of
Kristina Walter [30]).
features that are deemed valuable for therapy, until a sufficiently rich environment was created. Our communication
with sandtray experts was the main guide in determining the
most worthwhile features. A secondary concern was the cost
of a feature in terms of interface and interaction complexity.
DC1: maintain narrative potential — without characters,
there can be no story; thus, the use of figurines to represent
characters and objects is essential (Figure 2). It must be
possible to add figurines when a new character or object is
introduced, to move them around as the story progresses, and
to remove them when their part is over. Stacking objects,
such as a balanced stack of rocks or an animal on a house,
also has significant psychological connotations.
“[T]he temporary and unfixed nature of the sand-tray
pieces invites play and therefore the creation of narrative. Most other media result in a ‘snapshot’ in which
the narrative is implied but not played out.”
–T2
DC2: maintain the sandtray’s characteristic as primarily an
associative medium — with open media, such as paint and
clay, the artist (client) has a sense of the meaning as coming
from and being expressive of themselves. With a sandtray,
meaning is primarily associated with the objects, and usually
only one of the objects is the self representation. Thus, to
enable the development of associated meaning that allows
the client to express their particular story, a broad range of
objects or figurines is desirable. Some therapists group the
figurines they offer by category, which makes it easier to
find related figurines; others prefer a completely random presentation in which all figurines are mixed, which can trigger
more spontaneous associations. Although a digital system
could allow both options, we chose to use a random presentation to encourage free association. From a commercial
Also the temporary nature of the sand was described by the
therapists as being important. When a client sees a box of
sand, they immediately recognize that whatever they create
in that sandbox can be easily erased with a simple swipe of
the hand. A digital display is in this way similar to a physical
sandbox, because the pixels drawn on the screen are also in
some sense temporary. By clearing the screen or turning off
the monitor, whatever the client creates can be easily erased.
Recording is still possible, much like a sandtray session can
(and often is) videotaped.
DC3: exploring simple digital extensions — much of the design discussion with the therapists considered which aspects
of digital magic might make useful enhancements. In the
real world, it is not possible to instantly duplicate objects.
In the digital world, it is trivial. Being able to add multiple
copies of the same figurine allows for the creation of forests,
herds and families, with little or no cost in terms of interface
complexity. The sandtray therapists described this feature as
being particularly worthwhile.
Another real-world impossibility is resizing rigid objects
(again, digitally trivial). The size of an object has significant
psychological connotations: larger objects are perceived as
more important, more powerful or more menacing.
“. . . resizability could be a huge advantage of a virtual
play-table. It’s an ongoing issue that my toy collection
features a range of scales. I have some dinosaurs that
are smaller than my cockroaches. Children adapt and
play with this, and it sometimes suggests interesting
possibilities (like a giant baby who rescues a mom from
a tiny car), but scalability would give you the best of all
possible worlds. . . It’s such a rich metaphor: the sense
that in our psychological/creative world things do not
have ‘realistic’ sizes, they have metaphorical sizes.”–T2
DC4: use multi-touch to facilitate interpretation in the
therapist-client collaboration — The interactions between
the client and the sandtray are the focus of the therapist’s
observations, and thus a key aspect of this collaboration
is the awareness by the therapist of these interactions.
Furthermore, sandtray therapy is sometimes used for couples
or families, who will cooperatively act out a story, and even
a single client can ask the therapist to participate. We must
also design for such multi-person scenarios.
(a)
(b)
(c)
(d)
In the next three sections, we describe in detail how we
realize our design considerations in our implementation of
the virtual sandtray. We adapt existing tabletop display
techniques and technology to provide rich interactions so
the therapist can easily observe and infer information about
the client’s psyche from their interaction. Design principles
(DC1–DC4) are indicated for each implementation detail.
Implementation Details
We implemented our prototype using the SMART Table1 ,
which was specifically designed for children. Its small form
factor ensures that all corners of the table can be reached
by a child, and its sturdy design makes it suitable for use
in a practical setting. Moreover, it uses frustrated total internal reflection (FTIR) [11] and can detect up to 40 simultaneous touches, enabling interaction through multiple fingers for multiple people at once. The direct nature of multitouch technology supports awareness by the therapist of the
client’s interactions (DC4). These factors make the SMART
Table an ideal choice of hardware.
“. . . the principle is that a sand-tray should be just big
enough to fill the field of vision. This gives the sense
of an immersive world without requiring that the user
look around to take the whole thing in. . . My sandtray
is about the size that you suggest (75 by 52 cm), and
that feels about right.”
–T2
To enhance both the feeling of realism, and the narrative
abilities (DC1), we employ NVIDIA’s PhysX physics simulation engine2 in a similar way to Wilson et al. [31]. This
allows clients to make figurines fall down, roll around, knock
each other over, and to toss them around without any extra
interaction techniques or development effort.
Figurine Manipulation
The combination of precise control over the object being
moved, together with the physical reaction of the remainder
of the scene, provides the client with a platform for rich
expression through their narrative (DC1). With this system,
the physical movements of the client have a direct correlation with the response by the system, allowing the physical
movements to be interpreted by the therapist.
“. . . most children will depict battles at some point. Different varieties of aggression may be coming out in
this. . . ”
–T2
“positioning objects—includes orientation and ability to
push into sand.”
–T2
1
2
SMART Technologies. http://www.smarttech.com/
NVIDIA Corp. http://www.nvidia.com/object/nvidia physx.html
Figure 3: The first finger (a) moves the figurine in the plane. The second
finger rotates (b) about and moves (c) the figurine along the z axis. The
third finger (d) rotates the figurine about the x and y axes.
“moving objects—sometimes includes momentum, especially when throwing objects and lifting/dropping
them.”
–T2
“*stacking objects* (small objects placed on top of
larger ones, balanced stacks of rocks, animals in trees
or on houses etc.)”
–T2
“children love containment, frequently putting things
under or within other things. Another powerful metaphor”
–T2
To support this rich narrative, the client should be able to
freely move and rotate objects around on the surface, but
also to stack them (implying vertical movement). Thus, an
interaction technique is needed that provides the full six degrees of freedom (three translation and three rotation axes).
We use sticky fingers and opposable thumbs [14], which
is an extension of the three-touch technique by Hancock et
al. [12] and the proxy technique by Wilson et al. [31]. This
technique uses the order of contacts (assigned per object to
allow multi-object and multi-person interaction) to provide
all degrees of freedom: the first touch provides translation
in 2D (Figure 3a), the second touch provides rotation about
z (Figure 3b) and movement along the z-axis (Figure 3c),
and the third touch provides rotation about the x- and y-axes
(Figure 3d). The technique also uses ‘crossing’ (for each
contact) to allow acquisition of small objects.
Many figurines, such as human figures, will often be used in
a standing position. Because it is difficult or impossible to
make figurines with a small base stand upright, we add invisible pedestals at the bottom of these objects. The pedestals
are configured to collide only with the ground, and thus do
not cause unexpected interactions with other figurines.
Expressiveness Through Physics
One of the key advantages of the described interaction technique is that it offers precise control over an object, further
enabling the narrative (DC1). The use of a physics simulation, on the other hand, implies a certain imprecision and
Figurine Drawer. To allow the client to choose figurines
for their story (DC2), the collection of figurines needs to be
shown on the screen.
“. . . what really matters is that you’re always looking
at the toys themselves rather than some kind of menu
system. So whether you go with scrolling, zooming or a
‘drawer’ metaphor, what is presented to the user should
always be an image of some kind of toy or another.
Otherwise I think you’d be inviting a really cognitive
engagement with the system, which seems contrary to
the unique genius of the thing to me.”
–T2
Figure 4: The ring-shaped figurine drawer.
lack of control. To get the best of both worlds, during interaction the object is put into a ‘kinematic’ state, essentially giving it infinite mass. Thus, the object is controlled only by the
fingers and does not respond to forces in the physics engine,
but all other objects are controlled by the physics engine and
keep responding to the manipulated object. For example, it
is possible to drag a figurine around, knocking over other
figurines in its path, without losing control over the dragged
object. This combination allows the therapist to interpret
both the intended action on the object being controlled and
the physical reaction of other objects. When an object is
released, it retains the linear and angular velocity that it had
in the previous animation frame. This allows objects to be
tossed by moving them quickly, then releasing them. With
some practice it is also possible to make an object spin or fly
upwards, but this requires releasing two fingers within the
same animation frame. Objects can be stopped by simply
touching them. The ‘crossing’ feature also enables actions
such as sweeping across the surface with the side of the
hand. Although the actual interpretation of this gesture is
very different from the proxy objects introduced by Wilson
et al. [31], the net effect of objects being moved and pushing
other objects ahead of them is similar.
By enabling these familiar physical interactions, we provide
the client with a language for communicating to the therapist
(perhaps subconsciously) through the virtual objects themselves (DC1). The therapist can then interpret what actions
such as knocking over, tossing, and sweeping objects might
mean about the client’s psyche (DC4).
Drawers and Tools
To enable some of the digital extensions (DC3) and to support the accessibility of a wide range of figurines (DC2),
we introduce drawers and tools. Each of these tools was
designed to respond to the client’s touches with the same interaction technique as the figurines, using the physics engine
to impose constraints. We also designed the drawers to be
able to slide in and out of view using a handle to save screen
real estate. The drawers themselves can be tossed to quickly
open and close them, and figurines will bounce around them
in a natural way. Three such drawers are available: one
for selecting figurines, one for resizing them, and one for
painting on the sandtray floor.
Each figurine is shown as it would appear in the sandtray.
However, the tabletop screen is not large enough to show all
figurines at the same time with sufficient detail. We used a
ring-shaped drawer similar to a ‘Lazy Susan’, only a small
portion of which is visible at any given time (Figure 4).
This allows an unlimited number of figurines to be accessed.
When a figurine is removed from the drawer by dragging
(using the sticky fingers and opposable thumbs technique), a
copy will remain behind, allowing for quick and intuitive
duplication. Figurines can be removed from the scene
simply by putting them back into the drawer. This wide
range of figurines is beneficial for therapist’s observations,
as the client has more choice about which figurines to
pick. Therapists can also observe the client’s browsing and
decision-making processes, potentially involving inclusion,
exclusion, and/or copying of figurines.
Resizing Drawer. To provide the client with the (physically
impossible) ability to resize rigid objects (DC3), we provide
a resize drawer. An alternative way to implement this is the
two-finger ‘pinch’ gesture, in which two points on the object
are pulled apart or pushed together to grow or shrink the
object, similar to zooming. However, this gesture is already
mapped to vertical movement of the figurine. The use of
buttons or handles on the object would be harmful to the
sense of physical realism and might be easy for the therapist
to miss and therefore impede interpretation.
Instead, a drawer was added to the right side of the screen
which acts as a ‘resizing box’. One or more figurines can
be put into the drawer. A dial is provided on the side of the
drawer, with a ridged surface to suggest the ability to turn
it. The dial can be turned using a single finger. When it is
turned to the left, the figurines in the drawer shrink; when it
is turned right, they grow (Figure 5). A minimum and maximum size are provided to prevent figures from shrinking out
of sight or from becoming bigger than the display.
Painting Drawer. To provide the client with an open medium
that can be shaped at will (DC2), we provide them with the
ability to paint the background with different textures. On
first thought, the interaction to ‘paint’ on a touch sensitive
display could simply be like fingerpainting: wherever the surface is touched, the chosen paint colour appears. However,
combining this technique with the other interactions in the
virtual sandtray would require a mode switch, which would
likely be problematic [27], and particularly when multiple
people are using the system (DC4).
of our research was to support the ability for therapists to
recognize and interpret the actions of the client in a way
that was valuable for understanding more about the child’s
psyche. We designed this face-to-face session to address the
following questions:
• Can the therapists interpret the actions of a person using
the virtual sandtray and use them to gain insight about
their psyche?
Figure 5: The resizing drawer in three different stages. Upon spinning
the dial, the figurines inside the drawer will smoothly grow or shrink.
• How can the design be improved to better enable the therapists to gain insight about a client’s psyche?
Due to the fact that these therapists were typically distance
collaborators, we also took this opportunity to share domain
knowledge. While we had been iteratively discussing the
prototype design remotely, this was the first face-to-face opportunity to do live demonstrations by both groups.
Activities
In this day-long session, the sandtray therapists participated
in several activities: the therapists gave a demonstration of
sandtray therapy to the designers using traditional physical
figurines, the designers gave a demonstration of the virtual
sandtray prototype, the therapists conducted a mock therapy
session with the digital prototype, and all participated in a
follow-up interview/brainstorming session.
Figure 6: The painting system. In the drawer on the left, four buckets
of paint can be seen. The hose connects the bottom of ‘water paint’
bucket to the nozzle, from which the water texture flows.
The painting ability is provided through the use of a spray
nozzle tool. This nozzle can paint a texture on the floor
(Figure 6). When the nozzle is picked up, it will rotate to
point at the sandtray floor, which is the orientation in which
it will normally be used. When the nozzle is lifted up, away
from the floor, the region that is painted becomes larger.
The other part of the painting system is a drawer containing
buckets of paint. A hose running from the drawer to the
nozzle serves as a visual cue that they are related. A bucket
can be selected by touching it. As long as the bucket is
touched and the hose nozzle is in use, the texture paint it
contains will flow through the hose and out the nozzle. Usually, the dominant hand is used to move the nozzle, while
the non-dominant hand controls the paint selection. Paint
will flow from the nozzle when both the nozzle and a paint
bucket are being touched simultaneously. The nozzle can be
moved around without painting by releasing the bucket.
Both the resize drawer and the painting drawer involve very
explicit actions on the part of the client. The system’s reaction to these actions (i. e., through the physics engine) makes
the consequence of those actions available for interpretation
by the therapist (DC4).
PHASE III: FACE-TO-FACE FEEDBACK SESSION
To validate and iterate on our design, we performed a daylong session with two sandtray therapists together with three
of the designers. Neither therapist had any previous experience with digital tables. As previously stated, the focus
Physical Demonstration. We began with a demonstration
and instruction session from the perspective of the sandtray
therapists. In this part, the sandtray therapists described how
they performed sandtray therapy using physical figurines
and instructed the designers about the theory, logistics,
purpose, and essential components of the process of sandtray
therapy. We asked questions whenever something was
unclear and took written notes.
Prototype Demonstration. The designers provided an indepth demonstration of the digital sandtray prototype. We
spent approximately one hour explaining the details of how
to interact with our prototype and allowed both therapists
to experience using the system. We discussed our design
decisions, as well as several viable design alternatives.
The therapists were encouraged to share their thoughts and
comments. This part was videotaped.
Mock Therapy. Following the demonstration and a lunch
break, with the warning from the therapists that “some therapy might happen”, a designer played the role of the client
in a mock therapy session with a therapist, as the other three
participants (one therapist and two designers) observed. This
session lasted about an hour and was also videotaped.
Interview and Brainstorming. For the remainder of the day
(around 4 hours, including breaks), the two therapists and
three designers participated in an interview and brainstorming session. The interview began with several planned questions, and continued with an unstructured discussion of the
benefits, limitations, and next steps for future designs of
digital sandtray therapy. Designers again took written notes
and this part was not videotaped.
RESULTS & DISCUSSION
Of central import to the sandtray therapists, throughout all
three phases, was the issue of understanding the client’s psyche. An essential component of the exercise of sandtray
therapy is to give the therapist insight into what the client
is thinking and feeling through how they interact with objects. While sandtray therapists typically use physical props
to gain this insight, our design revealed that this process is
also possible with virtual ones. Instead of simply being a
digital representation of some underlying data or model, the
virtual objects in our system can take on symbolic meaning
in the same way that physical ones do, to the level of providing access into the mind of someone interacting with them.
In this section, we first describe what our research revealed
to be the essential components of interaction with virtual
objects. These essential components allow virtual objects
to cross the boundary from a digital representation to something that can allow the therapist access to the client’s psyche. We then describe specific design refinements of our
prototype that could address some of the issues that arose
in our iterative process. Note that in this section we are
discussing the results across all three phases of our design.
Use of Virtual Objects
There were several repeating themes about how a client’s
interaction with objects can help the therapist gain some insight about what they are thinking or feeling. These themes
include construction, storytelling, actions, and arrangement.
Construction. The therapists frequently identified the ability
to construct as an important aspect of the client’s interaction
with the sandtray. Construction is important because it “stimulates imagination”, and in stimulating imagination, the therapist can better access the client’s psyche. Several different
forms of construction became apparent in our sessions.
The first and most obvious form of construction was the
construction of the scene itself. The ability to freely move
and rotate the figurines made it possible to create a scene
that was composed of many different parts. The ability to
make multiple copies of a specific figurine played a key role
in this construction. For example, the client in our mock
therapy session placed four palm tree objects of different
sizes around an oasis. The therapist noted that the number
and size of these palm trees matched the number of members
in the client’s family. Furthermore, the therapist suggested
that this oasis may suggest that the family made the client
feel safe. This example demonstrates that this form of construction, made possible by our prototype, was a sufficiently
rich interaction for the therapist to gain insight.
Another key form of construction that was described by the
therapists was the ability to create barriers. This form of
construction was not directly supported by our prototype and
implies that another level of granularity (besides that of the
figurines) might be appropriate, where the virtual objects
that the client can interact with can be bent, folded, or attached to one another, like a fence or bricks.
Storytelling. Another important aspect of the sandtray interaction was the ability for both the therapist and client to
“tell a story”. This storytelling process might be brought
about in a variety of ways. For instance, the client might be
encouraged to just play in the sandtray, and then the therapist
might ask the client to explain the scene or elaborate on a
specific object and discuss what it means to them.
A key aspect of this storytelling is that the objects in the
scene can take on a variety of meanings. On the one hand,
an airplane can represent just an airplane (i. e., itself), but on
the other hand, it could represent a more abstract idea in the
client’s mind, such as flight or a desire to escape from something. The therapist’s experience with the virtual sandtray
prototype led them to state that they could easily interpret
these different meanings from the virtual objects.
Actions. The therapists also discussed several ways in which
the actions that the client uses to interact with the objects
can be key to gaining insight about their psyche. The most
commonly mentioned action from physical sandtray therapy
was the idea of burying an object. This burying process can
vary in meaning from client to client; it can imply things that
the client wishes to hide or keep secret, or it can be a sign
of aggression (e. g., suffocating). The ability for the therapist
to observe this burying process is fundamental to the therapy
session and was described as important to the therapist-client
collaboration for which we are designing. Our current prototype does not allow for this burying interaction, as our choice
of physics engine does not provide this functionality.
Other actions that may be relevant to the therapy were made
available through the combination of our interaction technique with the physics engine. In particular, the ability to
knock over one object with another, the ability to place an
object inside another, and the ability to toss an object across
the screen or drop an object from high above are actions that
a client may do and can help the therapist to understand what
is going on in the client’s mind when they create a scene.
Arrangement. The arrangement of objects in the scene was
also described as being highly important to the therapy process. In using our prototype, the therapists felt confident that
clients would be able to easily and freely arrange objects in a
way that would be useful for a therapy session. Although we
did not perform a formal evaluation of usability, we interpret
this confidence together with previous formal studies [12,31]
as a sign that the interaction technique that we included in
our prototype was sufficient for the type of object arrangements that the sandtray therapy requires. Furthermore, the
use of gravity through the physics engine and pedestals on
the base of each figurine facilitated this arrangement process.
Design Refinements
In addition to the high-level results regarding the efficacy
of virtual objects as meaning-carrying artifacts, we gathered
feedback about our current design that will help to inform future iterations. The therapists’ feedback suggests that these
improvements would facilitate interpreting a client’s actions.
Some words used by the therapists to describe the virtual
sandtray prototype were “relaxing and pleasurable”, “attractive”, and “appealing”. Both therapists commented that, in
its current form, the prototype might perhaps already be usable for therapy. On the other hand, comments were made
about the lack of sensory feedback: touch, sound, and even
smell. However, the application was described as being “still
quite tactile”. An interesting point raised was that the virtual
sandtray does not so much invite storytelling, but rather the
construction of a static scene. In that light, it might be more
related to art therapy, for example the making of a collage.
Figurines. A class of objects that was missing were arbitrary
objects that could be brought in by the therapist or the client
and play a more metaphorical role. A small cardboard box
could serve as a house, a stick could be used as a sword, or a
pine cone could represent a baby, covered by a handkerchief
to represent a blanket. The inability to bring such objects
in could be found to be limiting if many therapy sessions
are performed with the same, limited collection of figurines.
It was suggested that a possibility should be added to draw
or otherwise create one’s own figurines, but this would be
difficult to implement in an intuitive fashion, but might be
possible through Teddy [15] or ShapeShop [25]. An alternative would be to use a device such as the Microsoft Surface3
capable of recognizing physical objects and have the physical props interact with the virtual ones in our prototype.
The presentation of figurines in the drawer was identified
as potentially problematic; once a figurine was selected, the
lack of structure made it difficult to find related figurines.
Although the therapists who participated in the feedback session (T1 & T3) normally present the figurines in their therapy in an organized way (in contrast to T2), they did comment that the lack of ordering in the virtual sandtray prototype “stimulates more random aspects of the psyche”. This
difference in approach suggests that we should design for
some choice on the part of the therapist in how the figurines
are presented. We could add a way to configure whether the
objects are randomly sorted or organized in some fashion.
Alternatively, a hybrid approach could start with a random
presentation, but allow the client to easily find related figurines once a few have been selected.
Vertical Movement. Several problems were noted in relation to vertical movement of figurines. First, with the current top-down projection, it is not clear that the object is
actually moving up or down, instead of simply changing
size. This confusion was strongly reflected in the terminology used while discussing this action; even though people
know that the object is actually moving up and down in
the scene, they often still talk about “making it bigger” and
“making it smaller”. This suggests a disconnect between
the client’s actions and the therapist’s interpretation. This
might partly be blamed on the fact that the shadow is cast
directly below each object, which often causes the shadow
to be partly or completely obscured. A second shadow, cast
from the side, might improve interpretation of movement in
z; a projection that is not strictly top-down could also help.
A second problem is that the ‘sticky fingers’ paradigm [14]
implicitly makes lifting an object very sensitive. Especially
3
Microsoft Surface. http://www.microsoft.com/surface
when the two fingers start close together, a small movement
of the fingers will result in a large vertical motion. Doubling
the distance between the fingers will move the object twice
as close to the virtual camera, which is quite a large distance.
Perhaps it is better to let go of the stickiness of one of the
fingers. A more formal user study may be necessary to objectively determine which is better.
A third problem is that it is possible to move a figurine so
that it becomes invisible. For example, a figurine can be
pushed down right through a drawer from above, causing the
figurine to become hidden underneath the drawer. While this
‘feature’ may be one way of achieving the burying action
requested by the therapists, it may also be an unintended
consequence of the client’s actions, introducing ambiguity
in how the therapist interprets the action. It might be better
to keep a figurine always visible, by forcing it to be always
above everything else or by using digital effects such as outlines, shadows or transparency.
BEYOND THE DIGITAL SANDTRAY
On the one hand, our digital sandtray prototype is a single
point in the design space of interactive tabletops. Thus, our
iterative process may not yield results that generalize beyond
this design. However, our method is an example of how
we would recommend designing future interactive tabletop
systems and we would argue that many of the design decisions that we made can be adopted on other tabletop systems.
Furthermore, our feedback session provides some of the first
available evidence that virtual objects can be used in a real
application in a very different way than they are typically
used on desktop computers. In particular, the therapists in
our study pointed out the following aspects of our system as
particularly compelling:
Communication through virtual objects. The therapists
stated that the 3D interaction (sticky fingers and opposable
thumbs [14]) would be sufficiently rich for therapeutic
purposes. Specifically, the therapists felt they would be
able to gain insight about a client’s psyche based on his or
her interactions with a virtual object.
Deployable system. Our system is an example of an application that has been identified as usable as-is by the therapists.
Our method of designing and developing this application in
collaboration with sandtray therapists could be adopted for
tabletop systems in other domains. Specifically, our work
serves as a case study that cooperative design [9] may lead
to successful interactive tabletop systems.
Repurposeable virtual objects. The therapists felt that the
interaction with many individual 3D objects on an interactive table was both meaningful and usable. The therapists
illustrated this during mock therapy showing how sometimes
virtual objects were recognized as themselves (e. g., a rock
is a rock), as well as metaphors, symbols, or archetypes (i. e.,
representations of things from a person’s mind—for instance
a rock could represent a member of one’s family). This indicated that, in our system, virtual objects could be interpreted
by an observer as more than just digital representations of
data or information.
CONCLUSION
In this paper, we have presented the viability of a virtual
sandtray that has been identified as usable for therapy by
domain experts. Beyond the specific domain of art therapy,
this work also serves to inform the design process for tabletop display systems and provides some insight into how interaction with 3D objects on a table can be made useful in
practice. Specifically, the use of precise interaction and a
physics engine can together provide a richness that is sufficient for therapists to understand things about a client’s
psyche through their interactions with the virtual artifacts.
These artifacts thus can take on meaning in a way that is not
typically sought after in the design of traditional computer
applications. Future designers of tabletop systems can use
this work to inform how they can achieve similar levels of
rich interaction, and therefore make a new type of interface
between humans and virtual objects possible.
ACKNOWLEDGEMENTS
We would like to thank Natural Science and Engineering
Research Council of Canada, Albertas Informatics Circle
of Research Excellence, Alberta Ingenuity, and the Canadian Foundation of Innovation for research support. We also
thank the reviewers and members of the iLab for their helpful comments on this work.
REFERENCES
1. Agarawala, A., Balakrishnan, R. Keepin’ it real: pushing the
desktop metaphor with physics, piles and the pen. In Proc.
CHI, ACM (2006), 1283–1292.
2. Bers, M. U., Ackermann, E., Cassell, J., Donegan, B.,
Gonzalez-Heydrich, J., DeMaso, D. R., Strohecker, C., Lualdi,
S., Bromley, D., Karlin, J. Interactive storytelling
environments: coping with cardiac illness at Boston’s
Children’s Hospital. In Proc. CHI, ACM (1998), 603–610.
3. Bradway, K., McCoard, B. Sandplay: Silent Workshop of the
Psyche. Routledge, 1997.
4. Cao, X., Wilson, A. D., Balakrishnan, R., Hinckley, K.,
Hudson, S. E. ShapeTouch: leveraging contact shape on
interactive surfaces. In Proc. Tabletop, IEEE (2008), 129–136.
12. Hancock, M., Carpendale, S., Cockburn, A. Shallow-depth
3D interaction: Design and evaluation of one-, two- and
three-touch techniques. In Proc. CHI, ACM (2007),
1147–1156.
13. Hancock, M., Carpendale, S., Vernier, F. D., Wigdor, D., Shen,
C. Rotation and translation mechanisms for tabletop
interaction. In Proc. Tabletop, IEEE (2006), 79–88.
14. Hancock, M., ten Cate, T., Carpendale, S. Sticky tools: Full
6DOF force-based interaction for multi-touch tables. In Proc.
ITS, ACM (2009), 145–152.
15. Igarashi, T., Matsuoka, S., Tanaka, H. Teddy: a sketching
interface for 3D freeform design. In Proc. SIGGRAPH,
ACM/Addison-Wesley (1999), 409–416.
16. Kalff, D. M. Sandplay: A Psychotherapeutic Approach to the
Psyche. Sigo Press, 1980.
17. Kruger, R., Carpendale, S., Scott, S. D., Greenberg, S. Roles
of orientation in tabletop collaboration: comprehension,
coordination and communication. J. CSCW 13, 5–6 (2004),
501–537.
18. Kruger, R., Carpendale, S., Scott, S. D., Tang, A. Fluid
integration of rotation and translation. In Proc. CHI, ACM
(2005), 601–610.
19. Krüger, W., Fröhlich, B. The Responsive Workbench. CG&A
14, 3 (1994), 12–15.
20. Li, Y., Fontijn, W., Markopoulos, P. A tangible tabletop game
supporting therapy of children with cerebral palsy. In Proc.
Fun and Games, Springer-Verlag (2008), 182–193.
21. Liu, J., Pinelle, D., Sallam, S., Subramanian, S., Gutwin, C.
TNT: improved rotation and translation on digital tables. In
Proc. GI, Canadian Information Processing Society (2006),
25–32.
22. Morris, M. R., Cassanego, A., Paepcke, A., Winograd, T.,
Piper, A. M., Huang, A. Mediating group dynamics through
tabletop interface design. CG&A 26, 5 (2006), 65–73.
23. Piper, A. M., Hollan, J. D. Supporting medical conversations
between deaf and hearing individuals with tabletop displays.
In Proc. CSCW, ACM (2008), 147–156.
24. Ryokai, K., Cassell, J. StoryMat: A play space for
collaborative storytelling. In Proc. CHI, ACM (1999),
272–273.
5. Carpendale, M. Essence and Praxis in the Art Therapy Studio.
Trafford Publishing, 2009.
25. Schmidt, R., Wyvill, B., Sousa, M. C., Jorge, J. A. ShapeShop:
Sketch-based solid modeling with BlobTrees. In 2nd
Eurographics Workshop on Sketch-Based Interfaces and
Modeling (2005), 53–62.
6. Cassell, J., Ryokai, K. Making space for voice: technologies
to support children’s fantasy and storytelling. PUC 5, 3
(2001), 169–190.
26. Schuler, D., Namioka, A. (Eds.). Participatory Design:
Principles and Practices. L. Erlbaum Associates Inc., 1993.
7. Cunningham, J. (Ed.). Journal of Sandplay Therapy,
volume 18, number 2. Sandplay Therapists of America, 2009.
27. Sellen, A. J., Kurtenbach, G. P., Buxton, W. A. S. The
prevention of mode errors through sensory feedback. HCI 7, 2
(1992), 141–164.
8. Dietz, P., Leigh, D. DiamondTouch: a multi-user touch
technology. In Proc. UIST, ACM (2001), 219–226.
9. Greenbaum, J. M., Kyng, M. (Eds.). Design at Work:
Cooperative Design of Computer Systems. L. Erlbaum
Associates Inc., 1991.
10. Grubert, J., Carpendale, S., Isenberg, T. Interactive
stroke-based NPR using hand postures on large displays. In
Eurographics Short Papers, Eurographics Association (2008),
279–282.
11. Han, J. Y. Low-cost multi-touch sensing through frustrated
total internal reflection. In Proc. UIST, ACM (2005),
115–118.
28. Tse, E., Shen, C., Greenberg, S., Forlines, C. How pairs
interact over a multimodal digital table. In Proc. CHI, ACM
(2007), 215–218.
29. Walter, K. http://en.wikipedia.org/wiki/File:Sandspiel1.jpg.
30. Walter, K.
http://en.wikipedia.org/wiki/File:Sandspiel Figuren2.jpg.
31. Wilson, A. D., Izadi, S., Hilliges, O., Garcia-Mendoza, A.,
Kirk, D. Bringing physics to the surface. In Proc. UIST, ACM
(2008), 67–76.
32. Wu, M., Balakrishnan, R. Multi-finger and whole hand
gestural interaction techniques for multi-user tabletop
displays. In Proc. UIST, ACM (2003), 193–202.